The Coupon Effect: Quantifying Private School Subsidy with the Texas Vouchers

When discussing the impact of school vouchers, the conversation often revolves around the total number of applications. But as Fort Worth parents, we need to ask a deeper question: Are these vouchers "saving" students from underperforming public schools, or are they primarily subsidizing families who have already chosen private education or homeschooling?

By looking beyond the raw application numbers and calculating the Delta between Total Applications and Public Applications, we can reveal the true story. We call this the Coupon Effect.

The Query: (Total Apps - Previous Public School Apps) / Total Apps

This metric identifies districts where the voucher is functioning more like a "tax coupon" for families already outside the public system.

A high percentage here strongly suggests the program isn't rescuing students from public schools, but rather subsidizing existing private choices. Below, we dive into the data to compare this Coupon Effect against a proxy for the district's median household income (the percentage of students not economically disadvantaged).

To answer this, we need to combine two main data sources:

We calculate the pct_not_econ_disadv metric to proxy median household income for the district. The higher this percentage, the generally wealthier the district.

Before looking at the broader trend, a couple of extreme outliers in our data deserve a spotlight: Roscoe Collegiate ISD and Harmony ISD (not the charter school network).

When you join the TAPR demographic data with the voucher applications, these districts exhibit strange and out-sized behaviors. Below is a quick look at their raw numbers. Notice how massive the application volume is relative to the student population or typical application rates.

When we peel back the layers on these two extreme outliers, an incredible narrative emerges.

First, consider Roscoe Collegiate. This is a massive district serving over 12,000 students. Unfortunately, it struggles—carrying an overall "F" rating. For many families in this area, it is the only game in town. Yet despite the massive student body and the failing grade, only 93 families applied for a voucher. Here, the "Coupon Effect" is lower (~34%), indicating that a handful of brave souls are trying to exit a failing system, joining the small baseline of families already seeking alternative education.

Then, there is Harmony ISD. On the surface, things look perfectly fine—it holds a solid "B" rating. But look closer at the sheer volume of applications. Harmony is a tiny district with a total enrollment of just 895 students. Yet somehow, 320 applications were submitted for vouchers. We take away the 66 “coupon applications” from folks that were already in a private school or home school, and that is 254 Previous Public School applications. That means almost a third (28%) of the entire student body applied to leave a supposedly "good" school district. It suggests a staggering mass exodus and raises serious questions about what is happening beneath the surface of that "B" rating.

These exceptional outliers tell a fascinating story about how the voucher is being used at the fringes—either as a rare escape hatch in massive struggling districts, or as a floodgate in surprisingly small ones.

I'd be remiss if I didn't mention how we had to get this data. The original voucher application numbers were released in a deeply frustrating format: a static PDF with un-aligned columns, missing district IDs, and district names that didn't perfectly match the official TAPR reporting names.

In my opinion, publishing public data in a clickable PDF is as good as not releasing it at all. Whether it's a symptom of intentional obfuscation or sheer negligence in the talent appointed to the task, it's unacceptable.

We believe data belongs to the public, and you shouldn't have to be a data engineer to hold programs accountable. So, we've done the dirty work of parsing, cleaning, and joining the voucher applications with the official TAPR identifiers. You can download the cleaned dataset right here.

With the extreme outliers removed, the true picture emerges. If these vouchers were genuinely a "lifeline" specifically for lower-income families seeking to exit underperforming public schools, we would expect to see a strong diagonal relationship here—poorer districts (lower on the wealth proxy X-axis) utilizing the vouchers aggressively, with less utilization in wealthier districts.

But that's not what the graph shows.

Instead, we see a broad, horizontal cloud of points spanning the entire wealth spectrum. With a sort of ceiling of extreme points spanning the entire wealth spectrum as well. The "Coupon Effect" is uniformly high across the board, frequently sitting above 50% regardless of a district's median household income. This is powerful evidence that the voucher is functioning exactly as we suspected: primarily as a subsidy—a "tax coupon"—for a massive segment of families who were already opting out of the public school system anyway.

We can summarize this with a quick look at the Coupon Effect metric:

This author at FWSchools.org interprets this data as follows:

The Takeaway: The structural design of the voucher program allowed existing users to disproportionately apply for the vouchers. If the goal is to entice non-participants, the current distribution mechanism needs to be heavily revised to gatekeep existing users from claiming them.